Low impact crash detection for a vehicle
Systems and methods for detecting low impact collisions for a vehicle ( 100 ). The system includes at least one sensor ( 99, 110, 111, 115, 120 - 123, 125 - 136, 140, 141 ) and an electronic controller ( 150 ). The electronic controller ( 150 ) is configured to receive sensor data from the sensor ( 99, 110, 111, 115, 120 - 123, 125 - 136, 140, 141 ) and determine one or more features of the sensor data received from the at least one sensor. The electronic controller ( 150 ) is further configured to determine if a collision has occurred based upon the one or more features of the sensor data, and take at least one action in response to determining that the collision has occurred.
1. A system for detecting low impact collisions for a vehicle, the system comprising:
at least one sensor, and
an electronic controller configured to
receive sensor data from the sensor,
perform a plausibility step on the sensor data received from the at least one sensor to filter out unwanted data that could be misinterpreted as a low-impact collision to create filtered sensor data,
determine one or more features of the filtered sensor data,
determine if a collision has occurred based upon the one or more features of the filtered sensor data, and
take at least one action in response to determining that the collision has occurred.
2. The system of claim 1 , wherein the one or more features of the filtered sensor data include an energy from one or more spectrograms of the filtered sensor data.
3. The system of claim 1 , wherein the at least one sensor is one of a plurality of sensors, and wherein the plurality of sensors includes a plurality of peripheral contact sensors.
4. The system of claim 1 , wherein the at least one sensor is one of a plurality of sensors, and wherein the plurality of sensors includes a plurality of microphones.
5. The system of claim 1 , wherein the electronic controller is configured to determine if the collision has occurred using a machine learning algorithm.
6. The system of claim 5 , wherein the machine learning algorithm is a Bayesian classifier with a kernel function.
7. The system of claim 5 , wherein the machine learning algorithm is a neural network trained to detect the collision based upon the one or more features of the filtered sensor data.
8. The system of claim 1 , wherein the action is an action selected from the group consisting of outputting an indication of damage to a display and storing the filtered sensor data in a memory.
9. The system of claim 1 , wherein the vehicle is an autonomous vehicle, and wherein the action is an action selected from the group consisting of transmitting a notification of damage to a remote location via a wireless transceiver and transmitting a command to slow or stop the vehicle to a driving controller of the vehicle.
10. A method for detecting low-impact collisions for a vehicle, the method comprising
receiving, with an electronic controller, sensor data from at least one sensor,
performing, with the electronic controller, a plausibility step on the sensor data received from the at least one sensor to filter out unwanted data that could be misinterpreted as a low-impact collision to create filtered sensor data
determining, with the electronic controller, one or more features of the filtered sensor data,
determining, with the electronic controller, if a collision has occurred based upon the one or more features of the filtered sensor data, and
taking, with the electronic controller, at least one action in response to determining that the collision has occurred.
11. The method of claim 10 , wherein the one or more features of the filtered sensor data include an energy from one or more spectrograms of the filtered sensor data.
12. The method of claim 10 , wherein the at least one sensor is one of a plurality of sensors, and wherein the plurality of sensors includes a plurality of peripheral contact sensors.
13. The method of claim 10 , wherein the at least one sensor is one of a plurality of sensors, and wherein the plurality of sensors includes a plurality of microphones.
14. The method of claim 10 , further comprising determining, with the electronic controller, if the collision has occurred using a machine learning algorithm.
15. The method of claim 14 , wherein the machine learning algorithm is a Bayesian classifier with a kernel function.
16. The method of claim 14 , wherein the machine learning algorithm is a neural network trained to detect the collision based upon the one or more features of the filtered sensor data.
17. The method of claim 10 , wherein the action is an action selected from the group consisting of outputting, with the electronic controller, an indication of damage to a display and storing, with the electronic controller, the filtered sensor data in a memory.
18. The method of claim 10 , wherein the vehicle is an autonomous vehicle, and wherein the action is an action selected from the group consisting of transmitting, with the electronic controller, a notification of damage to a remote location via a wireless transceiver and transmitting, with the electronic controller, a command to slow or stop the vehicle to a driving controller of the vehicle.